hi, I am using some machine learning models. I want it to run with parallel processing. In ML forecast I have used num_threads. Still as per the CPU utilization and run time-its not doing multiprocessing. What else needs to be done to get multiprocessing in NIXTLA models? Thanks

which models are you using? num_threads parallelizes the feature engineering step, but most of the time is spent training the models

Thank you for your response. I am using XGboost,Random forest, Linear regression and LGBM.

Do you set n_jobs for those?

I used num_threads=-1.Should I use n_jobs?

You should've seen a warning about that then, num_threads should be a positive int, if you set -1 it gets set to 1

ohh thats why its not doing multiprocessing. I will set it to required number of cores. Thank you.